Medulloblastomas are highly invasive primitive neuroectodermal tumors of the cerebellum and the most common childhood malignant brain tumor, constituting 20-40% of pediatric brain tumors. Treating invasive intracranial brain tumors in children represents a significant challenge that is complicated further due to confined space and the need to preserve as much non-tumor, normal tissue as possible to avoid long-term cognitive dysfunction. In such cases, surgery is complicated and chemotherapy is prone to major side effects because cytotoxic drugs cannot differentially kill invading tumor cells surrounded by normal cells. In this EUREKA application, we present a highly innovative and unorthodox solution to this problem. We exploit the invasive nature of pediatric medulloblastomas by engineering a path of least resistance that moves tumors from the cerebellum to a pre-determined sub-dural location where they are killed. In this context, we introduce the term exvasion to mean the opposite of invasion - the tumor cells migrate and proliferate in a direction away from the primary tumor site, instead of invading deeper into the brain, and are thus directed to migrate to a safer , pre-determined sub-dural region to be killed. Our approach exploits the tumor s invasive character to move it away from the primary site by offering a path of least resistance specifically engineered to compete with its natural migratory pathway. Medulloblastoma migration and invasion along the leptomeningial pathway is facilitated by two elements: a) topographical cues presented by leptomeningial white matter tracts, and b) collagen rich extracellular matrix expressed along the leptomeningeal tract. Our design criteria to engineer a system to excavate tumors incorporates both of these elements; we propose to use aligned nanofiber-based polymeric thin films to mimic the topographical cues, and we coat these 10 micron-thin films with collagen I to mimic the ECM cues of the leptomeningial pathway. In addition to moving tumors out we propose to direct them to an engineered apoptosis-inducing hydrogel that will be implanted in a relatively safe sub-dural location. By directing tumor cell migration and invasion to an external sink, we will deliver tumor cells to the drug, rather than the current strategy of delivering the drug to the tumor, which is problematic due to the irregular vasculature and poor diffusivity of the tumor tissue. We have assembled a highly qualified, inter-disciplinary team consisting of a bioengineer/tumor drug delivery expert, Prof. Bellamkonda (PI), the Director of the Pediatric Oncology program at Emory School of Medicine and Children s Healthcare of Atlanta (CHOA), Prof. Macdonald, and a practicing pediatric neurosurgeon at Emory/CHOA who treats children with Medulloblastoma in his clinical practice, Prof. Brahma. We suggest that the proposed research is highly innovative, has the potential to open a new avenue for the treatment of solid tumors located intracranially, and represents significantly unorthodox research with a reasonably high chance of success from a highly qualified team worthy of EUREKA support.